POLAR: Ontology-Guided Risk Prevention for Tool-Calling LLM Agents

Key points

• LLM tool-use agents operate in dynamic environments where many actions carry operational risk.1
• Existing pre-emptive approaches either fine-tune the agent on chain-of-thought deliberation or compile natural-language guardrails into runtime checks, but they do so without exposing a structural, auditable verdict.1
• Evaluated on $τ^2$-bench across six agent models, POLAR improves mean task reward by 0.11 to 0.18 points on airline for four of six agents, but only eight of eighteen model--domain cells improve overall; retail and stronger agents often regress.1
• POLAR provides an auditable structural check and characterizes its task-utility trade-offs.1

Sources (1)

• [1]POLAR: Ontology-Guided Risk Prevention for Tool-Calling LLM Agents
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 6, 10:14 AM

  We propose POLAR, a guardrail framework for small tool-calling agents that assesses reversibility through a structured two-layer ontology.

  LLM tool-use agents operate in dynamic environments where many actions carry operational risk.

Extractive summary: sentences quoted from the sources.